Eigenvector-free solutions to \(AX = B\) with \(PX = XP\) and \(X^H = sX\) constraints (Q628932)

The matrix equation \(AX=B\) with \(PX=XP\) and \(X^H=sX\) is considered, where \(P\) is a given Hermitian involutory matrix and \(s=+1\). It is shown that by an eigenvalue decomposition of \(P\) one obtains a pair of similar problems which is equivalent with the initial one. Necessary and sufficient conditions for the existence of a solution are provided, as well as a representation of the general solution. Similar results are derived for eigenvalue-free solutions to the considered constrained problem. The optimal approximation problem \(\min\|X-E\|_F\) is studied, where \(E\in\mathbb{C}^{n\times n}\) is a given matrix and \(X_s\) is the solution set of the constrained problem. Eigenvalue-involved and eigenvalue-free solutions for the optimal problem are provided. A similar approach is presented for the extended problem of the matrix equations \(AX=B,\;XC=D,\;A,B\in\mathbb{C}^{m\times n},\;C,D\in\mathbb{C}^{n\times p}\), with the same constraints. Some numerical examples illustrate the effectiveness of the proposed methods.